Speech Recognition Software

There are a few products out there that can convert the spoken word to text, but the two standouts are still:

Dragon Naturally Speaking Preferred (currently version 9) for Windows, with an education price of $154 (+GST) and

MacSpeech Dictate for Mac (Intel based Macs only) for $363 (+GST) (which includes a headset).

Despite years of development, we are still a long way off a perfect solution, and both these require you to train the computer in your speech style and syntax. I have found in the past that the accuracy of the package is greatly improved if you submit some of your written work into it, as it then learns how you write, and can improve its ability to anticipate your next word (similar to how we interpret what someone is saying to us – from the context, we have an idea of what is coming up, so are listening out for specific words rather than the entire vocabulary).

Where the programs will still undergo significant development is understanding multiple speakers. At the moment, they are very good at recognising an individual’s spoken words, but cannot handle (for example) a group discussion.